Código/Articulo_La complejidad del sistema geoelectoral mexicano a nivel municipal

Análisis de Agrupamiento y Estratificacion

Autor: Miguel David Alvarez Hernández (mdalvarezh@gmail.com)

import datetime
now = datetime.datetime.now()
print ("Última versión:")
print (now.strftime("%Y-%m-%d %H:%M:%S"))

Última versión:
2021-08-10 23:01:05

Setup

import pandas as pd  
import numpy as np
import pandas_profiling
from itertools import combinations 
from dateutil.parser import parse 
import matplotlib as mpl
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import pyplot
from matplotlib.lines import Line2D
import seaborn as sns
import plotly.express as px

#parametros de las gráficas 3d
elev = 50.0
azim = 10.0

#se define diccionario de colores
color_dict = dict({'G0':'blue',
                  'G1':'green',
                  'G2': 'orange',
                  'G3': 'grey',
                  'G4': 'red',
                  'G5': 'magenta'})

#personalizar leyenda
legend_elements_6 = [Line2D([0], [0], marker='o', markerfacecolor='magenta', color='w', label='G5'),
                   Line2D([0], [0], marker='o', markerfacecolor='red', color='w', label='G4'),
                   Line2D([0], [0], marker='o', markerfacecolor='grey', color='w', label='G3'),
                   Line2D([0], [0], marker='o', markerfacecolor='orange', color='w', label='G2'),
                   Line2D([0], [0], marker='o', markerfacecolor='green', color='w', label='G1'),
                   Line2D([0], [0], marker='o', markerfacecolor='blue', color='w', label='G0')]

legend_elements_5 = [Line2D([0], [0], marker='o', markerfacecolor='red', color='w', label='G4'),
                   Line2D([0], [0], marker='o', markerfacecolor='grey', color='w', label='G3'),
                   Line2D([0], [0], marker='o', markerfacecolor='orange', color='w', label='G2'),
                   Line2D([0], [0], marker='o', markerfacecolor='green', color='w', label='G1'),
                   Line2D([0], [0], marker='o', markerfacecolor='blue', color='w', label='G0')]

legend_elements_4 = [Line2D([0], [0], marker='o', markerfacecolor='grey', color='w', label='G3'),
                   Line2D([0], [0], marker='o', markerfacecolor='orange', color='w', label='G2'),
                   Line2D([0], [0], marker='o', markerfacecolor='green', color='w', label='G1'),
                   Line2D([0], [0], marker='o', markerfacecolor='blue', color='w', label='G0')]

Importación de datos procesados

#datos procesados
datos = pd.read_csv("C:/Users/miguel.alvarez/Google Drive/PREPRINTS_PONENCIAS/[Articulo]_Complejidad sistema geo-electoral_municipios/Code_Mun/1_Procesamiento/Resultados_Datos-procesados_mun2019.csv", encoding='utf-8')

print(datos.dtypes)
datos

EDO                   int64
NOMBRE_ESTADO        object
MUN                   int64
NOMBRE_MUNICIPIO     object
Coef_Var_PE         float64
Coef_Var_LNE        float64
Var_Prop_PE         float64
Var_Prop_LNE        float64
Autocorr_PE         float64
Autocorr_LNE        float64
Razon_LNE_PE        float64
PE                  float64
LNE                 float64
TC_LNE_2019         float64
NUM_SECC_RUR          int64
Area_Km2            float64
Densidad_LNE        float64
dtype: object

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	Coef_Var_PE	Coef_Var_LNE	Var_Prop_PE	Var_Prop_LNE	Autocorr_PE	Autocorr_LNE	Razon_LNE_PE	PE	LNE	TC_LNE_2019	NUM_SECC_RUR	Area_Km2	Densidad_LNE
0	1	AGUASCALIENTES	10	SAN FRANCISCO DE LOS ROMO	0.009708	0.008903	0.020041	0.017733	0.988393	0.970914	0.989272	35518.0	34917.0	0.037498	3	9.791576	3566.024547
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	0.003117	0.002356	0.005584	0.008887	0.791791	0.951687	0.991803	7004.0	6914.0	0.000289	4	148.866104	46.444421
2	1	AGUASCALIENTES	9	TEPEZALA	0.003911	0.005070	0.007903	0.011921	0.969294	0.832575	0.990635	15733.0	15545.0	0.005108	6	6.271495	2478.675412
3	1	AGUASCALIENTES	11	EL LLANO	0.005733	0.004768	0.009983	0.014232	0.938868	0.962703	0.991016	15298.0	15067.0	-0.003373	6	10.006312	1505.749552
4	1	AGUASCALIENTES	2	ASIENTOS	0.005650	0.004824	0.009252	0.015397	0.993797	0.837235	0.989080	34722.0	34222.0	0.001874	11	23.049424	1484.722551
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2453	32	ZACATECAS	45	TEPECHITLAN	0.002327	0.001991	0.017748	0.021905	0.992283	0.577314	0.983523	7061.0	6959.0	0.066677	13	190.978234	36.438707
2454	32	ZACATECAS	14	GENERAL FRANCISCO R. MURGUIA	0.004739	0.002967	0.022931	0.026072	0.982714	0.911464	0.986450	17610.0	17372.0	0.082705	32	1190.248462	14.595272
2455	32	ZACATECAS	44	TABASCO	0.005701	0.005293	0.032716	0.035416	0.952155	0.928948	0.983221	12689.0	12527.0	0.100307	16	235.714183	53.144872
2456	32	ZACATECAS	9	CHALCHIHUITES	0.004951	0.003692	0.019649	0.024771	0.655389	0.819352	0.982861	8339.0	8222.0	0.075052	10	13.707054	599.837116
2457	32	ZACATECAS	11	TRINIDAD GARCIA DE LA CADENA	0.003147	0.002873	0.033702	0.041745	0.864769	-0.273110	0.972699	2853.0	2779.0	0.113381	3	175.709403	15.815887

2458 rows × 17 columns

#se reordena y eliminan del dataframe las variables que no se usarán
datos_mun = datos.drop(columns=['Coef_Var_PE', 'Var_Prop_PE', 'Autocorr_PE','Area_Km2'])
datos_mun = datos_mun.reindex(columns= ['EDO','NOMBRE_ESTADO','MUN','NOMBRE_MUNICIPIO','NUM_SECC_RUR','LNE','Densidad_LNE','Razon_LNE_PE','TC_LNE_2019','Coef_Var_LNE','Var_Prop_LNE','Autocorr_LNE'])

#verificar si hay valores nulos en datos_mun
datos_mun.isnull().sum()

EDO                 0
NOMBRE_ESTADO       0
MUN                 0
NOMBRE_MUNICIPIO    0
NUM_SECC_RUR        0
LNE                 0
Densidad_LNE        0
Razon_LNE_PE        0
TC_LNE_2019         0
Coef_Var_LNE        0
Var_Prop_LNE        0
Autocorr_LNE        2
dtype: int64

#se eliminan las filas (municipios) con valores vacíos
datos_mun_limpio = datos_mun.dropna()
datos_mun_limpio = datos_mun_limpio.reset_index(drop=True) #para reordenar el índice del nuevo dataframe
datos_mun_limpio.isnull().sum()

EDO                 0
NOMBRE_ESTADO       0
MUN                 0
NOMBRE_MUNICIPIO    0
NUM_SECC_RUR        0
LNE                 0
Densidad_LNE        0
Razon_LNE_PE        0
TC_LNE_2019         0
Coef_Var_LNE        0
Var_Prop_LNE        0
Autocorr_LNE        0
dtype: int64

#se transforman algunos indicadores para analizar escenarios
datos_mun_limpio_tf = datos_mun_limpio.copy()

#np-log de LNE, Densidad
datos_mun_limpio_tf['LNE'] = np.log(datos_mun_limpio_tf['LNE'])
datos_mun_limpio_tf['Densidad_LNE'] = np.log(datos_mun_limpio_tf['Densidad_LNE'])
#datos_mun_limpio_tf['NUM_SECC_RUR'] = np.log(datos_mun_limpio_tf['NUM_SECC_RUR']+1)

#se separan los indicadores de los metadatos 

#metadatos
metadatos = datos_mun_limpio[['EDO', 'NOMBRE_ESTADO', 'MUN', 'NOMBRE_MUNICIPIO']]

#8 indicadores (features)
datos_mun_limpio_f = datos_mun_limpio[['NUM_SECC_RUR','LNE', 'Densidad_LNE', 'Razon_LNE_PE', 'TC_LNE_2019', 'Coef_Var_LNE', 'Var_Prop_LNE','Autocorr_LNE']]

#8 indicadores con transformación log (features)
datos_mun_limpio_f_tf = datos_mun_limpio_tf[['NUM_SECC_RUR','LNE', 'Densidad_LNE', 'Razon_LNE_PE', 'TC_LNE_2019', 'Coef_Var_LNE', 'Var_Prop_LNE','Autocorr_LNE']]

#7 indicadores(features)
datos_mun_limpio_f_ind7 = datos_mun_limpio[['NUM_SECC_RUR','LNE', 'Densidad_LNE', 'Razon_LNE_PE', 'TC_LNE_2019', 'Coef_Var_LNE', 'Autocorr_LNE']]

#7 indicadores con transformación log (features)
datos_mun_limpio_f_tf_ind7 = datos_mun_limpio_tf[['NUM_SECC_RUR','LNE', 'Densidad_LNE', 'Razon_LNE_PE', 'TC_LNE_2019', 'Coef_Var_LNE', 'Autocorr_LNE']]

Resultados de los escenarios PCA

#escenario con 7 indicadores (sin Coef_Var_Prop)
df_pca_results_ind7 = pd.read_csv("C:/Users/miguel.alvarez/Google Drive/PREPRINTS_PONENCIAS/[Articulo]_Complejidad sistema geo-electoral_municipios/Code_Mun/2_Exploratorio-PCA/Resultados_Mun_PCA_ind7.csv", dtype={'EDO':int,'MUN':int})

#formamos un np-array con las columnas_pca para la sección de métodos de agrupamiento
columnas_7pca = ['CP1', 'CP2', 'CP3','CP4', 'CP5', 'CP6','CP7']
columnas_7pca_3 = ['CP1', 'CP2', 'CP3']
x_pca = df_pca_results_ind7.loc[:, columnas_7pca_3].values

print(x_pca.shape)

(2456, 3)

Definición de la prueba de Hopkins (https://www.kaggle.com/ashydv/country-clustering-hierarchical-clustering-pca)

from sklearn.neighbors import NearestNeighbors
from random import sample
from numpy.random import uniform
import numpy as np
from math import isnan
 
def hopkins(X):
    d = X.shape[1]
    #d = len(vars) # columns
    n = len(X) # rows
    m = int(0.1 * n) 
    nbrs = NearestNeighbors(n_neighbors=1).fit(X.values)
 
    rand_X = sample(range(0, n, 1), m)
 
    ujd = []
    wjd = []
    for j in range(0, m):
        u_dist, _ = nbrs.kneighbors(uniform(np.amin(X,axis=0),np.amax(X,axis=0),d).reshape(1, -1), 2, return_distance=True)
        ujd.append(u_dist[0][1])
        w_dist, _ = nbrs.kneighbors(X.iloc[rand_X[j]].values.reshape(1, -1), 2, return_distance=True)
        wjd.append(w_dist[0][1])
 
    H = sum(ujd) / (sum(ujd) + sum(wjd))
    if isnan(H):
        print(ujd, wjd)
        H = 0
 
    return H

Agrupamiento y construcción de la estratificación de complejidad electoral

(7 indicadores, 3 CP)

Analisis de hopkins

hopkins(pd.DataFrame(x_pca))

0.9754367867461262

Análisis de silueta, dendograma y configuración óptima de grupos con Ward

Cálculo de los silhouette_scores para diferentes números de clusters en ward

from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import silhouette_samples, silhouette_score
import matplotlib.cm as cm

#Datos en el espacio de las componentes principales
X=x_pca

#numero de clusters/grupos que serán analizados
range_n_clusters = [2, 3, 4, 5,6,7]

for n_clusters in range_n_clusters:
    # Create a subplot with 1 row and 2 columns
    fig, (ax1, ax2) = plt.subplots(1, 2)
    fig.set_size_inches(18, 7)

    # The 1st subplot is the silhouette plot
    # The silhouette coefficient can range from -1, 1 
    ax1.set_xlim([-1, 1])
    # The (n_clusters+1)*10 is for inserting blank space between silhouette
    # plots of individual clusters, to demarcate them clearly.
    ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])

    # Initialize the clusterer with n_clusters value and a random generator
    # seed of 0 for reproducibility.
    clusterer = AgglomerativeClustering(
        n_clusters=n_clusters, linkage='ward')
    cluster_labels = clusterer.fit_predict(X)

    # The silhouette_score gives the average value for all the samples.
    # This gives a perspective into the density and separation of the formed clusters
    silhouette_avg = silhouette_score(X, cluster_labels)
    print("For n_clusters =", n_clusters,
          "The average silhouette_score is :", silhouette_avg)

    # Compute the silhouette scores for each sample
    sample_silhouette_values = silhouette_samples(X, cluster_labels)

    y_lower = 10
    for i in range(n_clusters):
        # Aggregate the silhouette scores for samples belonging to
        # cluster i, and sort them
        ith_cluster_silhouette_values = \
            sample_silhouette_values[cluster_labels == i]

        ith_cluster_silhouette_values.sort()

        size_cluster_i = ith_cluster_silhouette_values.shape[0]
        y_upper = y_lower + size_cluster_i

        color = cm.nipy_spectral(float(i) / n_clusters)
        ax1.fill_betweenx(np.arange(y_lower, y_upper),
                          0, ith_cluster_silhouette_values,
                          facecolor=color, edgecolor=color, alpha=0.7)

        # Label the silhouette plots with their cluster numbers at the middle
        ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))

        # Compute the new y_lower for next plot
        y_lower = y_upper + 10  # 10 for the 0 samples

    ax1.set_title("Diagrama de silueta por cada grupo.")
    ax1.set_xlabel("Coeficiente de silueta")
    ax1.set_ylabel("Grupos")

    # The vertical line for average silhouette score of all the values
    ax1.axvline(x=silhouette_avg, color="red", linestyle="--")

    ax1.set_yticks([])  # Clear the yaxis labels / ticks

    # 2nd Plot showing the actual clusters formed
    colors = cm.nipy_spectral(cluster_labels.astype(float) / n_clusters)
    ax2.scatter(X[:, 0], X[:, 1], marker='.', s=30, lw=0, alpha=0.7,
                c=colors, edgecolor='k')

    ax2.set_title("Subespacio CP1-CP2")
    ax2.set_xlabel("CP1 (ICE principal)")
    ax2.set_ylabel("CP2 (segundo ICE)")

    plt.suptitle(("Análisis de silueta con Ward (7 indicadores base), "
                  "número de estratos = %d, " % n_clusters +
                 "score de silueta promedio = %1.3f" % silhouette_avg),
                 fontsize=14, fontweight='bold')
    #plt.savefig('Pyplot_silhouette_score_ind7_Ward_'+str(n_clusters)+'.png',dpi=600)

For n_clusters = 2 The average silhouette_score is : 0.7833096821167509
For n_clusters = 3 The average silhouette_score is : 0.26020124716848825
For n_clusters = 4 The average silhouette_score is : 0.2581498406179149
For n_clusters = 5 The average silhouette_score is : 0.28166477517791494
For n_clusters = 6 The average silhouette_score is : 0.2855594159970792
For n_clusters = 7 The average silhouette_score is : 0.22351196302289245

Dendrograma

import scipy.cluster.hierarchy as shc
plt.figure(figsize=(10,7))  
plt.title("Dendrograma, método de Ward")  
dend = shc.dendrogram(shc.linkage(X, method='ward'))
#plt.savefig('Pyplot_dendrograma_ind7_Ward.png',dpi=600)

Se elige la configuración de 4 clusters (es la tercera configuración con el mayor valor promedio de silueta, pero presenta un agrupamiento adecuado en la proyección bidimensional y en el dendrograma)

#se obtienen las etiquetas predichas por Ward para la configuración con el mayor
#valor promedio del silhouete score (n=5)
num_clusters_ind7 = 5
clusterer_K = AgglomerativeClustering(n_clusters=num_clusters_ind7, affinity='euclidean', linkage='ward')
clusterer_K_labels = clusterer_K.fit_predict(X)

#se obtienen las etiquetas predichas por ward
y_pred_ward = clusterer_K_labels
#se guarda el nparray en un dataframe
y_pred_ward = pd.DataFrame(y_pred_ward, columns=['Estrato'])

#se transforman las etiquetas numéricas a str
etiquetas_grupos = {0:'G0', 1:'G1', 2:'G2', 3:'G3', 4:'G4', 5:'G5'}
y_pred_ward['Estrato']=y_pred_ward['Estrato'].apply(lambda x: etiquetas_grupos[x])

#concatenamos los resultados de PCA con las etiquetas de ward
df_pca_results_ind7_ward = pd.concat([df_pca_results_ind7, y_pred_ward], axis = 1)

#se guardan los resultados en un csv
df_pca_results_ind7_ward.to_csv(r'Resultados_Mun_PCA_Ward_ind7.csv', index = None)

df_pca_results_ind7_ward

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	CP1	CP2	CP3	CP4	CP5	CP6	CP7	Estrato
0	1	AGUASCALIENTES	10	SAN FRANCISCO DE LOS ROMO	-0.290827	1.304370	-0.470542	1.249692	-0.985389	0.343395	-0.883019	G1
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	-0.215022	-0.976397	-0.423919	1.386989	-1.260267	-0.625396	0.304296	G2
2	1	AGUASCALIENTES	9	TEPEZALA	-0.597435	0.071978	-0.716279	1.371428	-0.870282	-0.309656	-0.647832	G1
3	1	AGUASCALIENTES	11	EL LLANO	-0.382682	-0.278937	-0.606074	1.675445	-0.996878	-0.462573	-0.289389	G2
4	1	AGUASCALIENTES	2	ASIENTOS	-0.428400	-0.232114	-0.320612	1.533638	-0.432850	-0.508456	-0.335265	G2
...	...	...	...	...	...	...	...	...	...	...	...	...
2451	32	ZACATECAS	45	TEPECHITLAN	0.336526	-0.414118	0.879856	-0.236011	-0.054547	-0.515434	0.024942	G0
2452	32	ZACATECAS	14	GENERAL FRANCISCO R. MURGUIA	-0.052455	-0.693651	2.394801	0.574858	0.201790	-0.068696	-0.577988	G0
2453	32	ZACATECAS	44	TABASCO	0.815847	0.275372	1.779525	0.018950	-0.257829	0.094995	-0.099322	G0
2454	32	ZACATECAS	9	CHALCHIHUITES	0.562309	0.233027	0.954854	-0.056961	-0.390447	-0.419723	-0.098396	G0
2455	32	ZACATECAS	11	TRINIDAD GARCIA DE LA CADENA	1.391414	0.517387	0.405960	-2.500965	1.204864	-0.454846	0.366341	G0

2456 rows × 12 columns

Análisis de silueta, dendograma y configuración óptima de grupos con complete linkage

Cálculo de los silhouette_scores para diferentes números de clusters en complete linkage

from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import silhouette_samples, silhouette_score
import matplotlib.cm as cm

#Datos en el espacio de las componentes principales
X=x_pca

#numero de clusters/grupos que serán analizados
range_n_clusters = [2, 3, 4, 5,6,7]

for n_clusters in range_n_clusters:
    # Create a subplot with 1 row and 2 columns
    fig, (ax1, ax2) = plt.subplots(1, 2)
    fig.set_size_inches(18, 7)

    # The 1st subplot is the silhouette plot
    # The silhouette coefficient can range from -1, 1 
    ax1.set_xlim([-1, 1])
    # The (n_clusters+1)*10 is for inserting blank space between silhouette
    # plots of individual clusters, to demarcate them clearly.
    ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])

    # Initialize the clusterer with n_clusters value and a random generator
    # seed of 0 for reproducibility.
    clusterer = AgglomerativeClustering(
        n_clusters=n_clusters, linkage='complete')
    cluster_labels = clusterer.fit_predict(X)

    # The silhouette_score gives the average value for all the samples.
    # This gives a perspective into the density and separation of the formed clusters
    silhouette_avg = silhouette_score(X, cluster_labels)
    print("For n_clusters =", n_clusters,
          "The average silhouette_score is :", silhouette_avg)

    # Compute the silhouette scores for each sample
    sample_silhouette_values = silhouette_samples(X, cluster_labels)

    y_lower = 10
    for i in range(n_clusters):
        # Aggregate the silhouette scores for samples belonging to
        # cluster i, and sort them
        ith_cluster_silhouette_values = \
            sample_silhouette_values[cluster_labels == i]

        ith_cluster_silhouette_values.sort()

        size_cluster_i = ith_cluster_silhouette_values.shape[0]
        y_upper = y_lower + size_cluster_i

        color = cm.nipy_spectral(float(i) / n_clusters)
        ax1.fill_betweenx(np.arange(y_lower, y_upper),
                          0, ith_cluster_silhouette_values,
                          facecolor=color, edgecolor=color, alpha=0.7)

        # Label the silhouette plots with their cluster numbers at the middle
        ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))

        # Compute the new y_lower for next plot
        y_lower = y_upper + 10  # 10 for the 0 samples

    ax1.set_title("Diagrama de silueta por cada grupo.")
    ax1.set_xlabel("Coeficiente de silueta")
    ax1.set_ylabel("Grupos")

    # The vertical line for average silhouette score of all the values
    ax1.axvline(x=silhouette_avg, color="red", linestyle="--")

    ax1.set_yticks([])  # Clear the yaxis labels / ticks

    # 2nd Plot showing the actual clusters formed
    colors = cm.nipy_spectral(cluster_labels.astype(float) / n_clusters)
    ax2.scatter(X[:, 0], X[:, 1], marker='.', s=30, lw=0, alpha=0.7,
                c=colors, edgecolor='k')

    ax2.set_title("Subespacio CP1-CP2")
    ax2.set_xlabel("CP1 (ICE principal)")
    ax2.set_ylabel("CP2 (segundo ICE)")

    plt.suptitle(("Análisis de silueta con Complete Linkage (7 indicadores base), "
                  "número de estratos = %d, " % n_clusters +
                 "score de silueta promedio = %1.3f" % silhouette_avg),
                 fontsize=14, fontweight='bold')
    #plt.savefig('Pyplot_silhouette_score_ind7_complete_'+str(n_clusters)+'.png',dpi=600)

For n_clusters = 2 The average silhouette_score is : 0.7427739664649179
For n_clusters = 3 The average silhouette_score is : 0.4489718251589875
For n_clusters = 4 The average silhouette_score is : 0.44635344146353956
For n_clusters = 5 The average silhouette_score is : 0.44323842615769177
For n_clusters = 6 The average silhouette_score is : 0.3761542262111524
For n_clusters = 7 The average silhouette_score is : 0.39650348186741535

Dendrograma

import scipy.cluster.hierarchy as shc
plt.figure(figsize=(10,7))  
plt.title("Dendrograma, 3 estratos, método de complete linkage")  
dend = shc.dendrogram(shc.linkage(X, method='complete'))
#plt.savefig('Pyplot_dendrograma_ind7_complete.png',dpi=600)

Se elige la configuración de 3 clusters

#se obtienen las etiquetas predichas por Ward para la configuración con el mayor
#valor promedio del silhouete score (n=5)
num_clusters_ind7 = 3
clusterer_K = AgglomerativeClustering(n_clusters=num_clusters_ind7, affinity='euclidean', linkage='complete')
clusterer_K_labels = clusterer_K.fit_predict(X)

#se obtienen las etiquetas predichas por ward
y_pred_ward = clusterer_K_labels
#se guarda el nparray en un dataframe
y_pred_ward = pd.DataFrame(y_pred_ward, columns=['Estrato'])

#se transforman las etiquetas numéricas a str
etiquetas_grupos = {0:'G0', 1:'G1', 2:'G2'}
y_pred_ward['Estrato']=y_pred_ward['Estrato'].apply(lambda x: etiquetas_grupos[x])

#concatenamos los resultados de PCA con las etiquetas de ward
df_pca_results_ind7_complete = pd.concat([df_pca_results_ind7, y_pred_ward], axis = 1)

#se guardan los resultados en un csv
df_pca_results_ind7_complete.to_csv(r'Resultados_Mun_PCA_complete_ind7.csv', index = None)

df_pca_results_ind7_complete

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	CP1	CP2	CP3	CP4	CP5	CP6	CP7	Estrato
0	1	AGUASCALIENTES	10	SAN FRANCISCO DE LOS ROMO	-1.061899	0.788043	-0.837586	-0.614461	-0.088128	-0.782747	-0.238578	G0
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	-3.242619	-1.163180	-0.391435	0.193575	0.880911	0.128676	-1.012632	G0
2	1	AGUASCALIENTES	9	TEPEZALA	-2.714649	-0.207140	-0.939179	0.184211	0.615393	-0.770295	-0.744057	G0
3	1	AGUASCALIENTES	11	EL LLANO	-2.489077	-0.324405	-0.411126	-0.029804	0.468563	-0.389211	-0.141379	G0
4	1	AGUASCALIENTES	2	ASIENTOS	-2.499466	-0.395191	-0.509621	0.718468	0.620966	-0.484175	-0.511286	G0
...	...	...	...	...	...	...	...	...	...	...	...	...
2453	32	ZACATECAS	45	TEPECHITLAN	0.548172	-0.308683	0.679952	0.209243	-0.452260	0.088731	-0.062292	G0
2454	32	ZACATECAS	14	GENERAL FRANCISCO R. MURGUIA	0.875699	0.306366	1.949664	0.996421	0.061302	-0.572234	-0.267121	G0
2455	32	ZACATECAS	44	TABASCO	1.864008	-0.162682	0.715202	0.470561	0.292022	-0.127472	-0.194333	G0
2456	32	ZACATECAS	9	CHALCHIHUITES	0.886651	-0.151959	0.337662	0.033250	-0.481215	-0.003173	-0.078652	G0
2457	32	ZACATECAS	11	TRINIDAD GARCIA DE LA CADENA	3.059160	-1.073151	-0.872196	0.842564	-0.248571	0.300834	-0.133138	G2

2458 rows × 12 columns

Análisis de silueta, dendograma y configuración óptima de grupos con average linkage

Cálculo de los silhouette_scores para diferentes números de clusters en average linkage

from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import silhouette_samples, silhouette_score
import matplotlib.cm as cm

#Datos en el espacio de las componentes principales
X=x_pca

#numero de clusters/grupos que serán analizados
range_n_clusters = [2, 3, 4, 5,6,7]

for n_clusters in range_n_clusters:
    # Create a subplot with 1 row and 2 columns
    fig, (ax1, ax2) = plt.subplots(1, 2)
    fig.set_size_inches(18, 7)

    # The 1st subplot is the silhouette plot
    # The silhouette coefficient can range from -1, 1 
    ax1.set_xlim([-1, 1])
    # The (n_clusters+1)*10 is for inserting blank space between silhouette
    # plots of individual clusters, to demarcate them clearly.
    ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])

    # Initialize the clusterer with n_clusters value and a random generator
    # seed of 0 for reproducibility.
    clusterer = AgglomerativeClustering(
        n_clusters=n_clusters, linkage='average')
    cluster_labels = clusterer.fit_predict(X)

    # The silhouette_score gives the average value for all the samples.
    # This gives a perspective into the density and separation of the formed clusters
    silhouette_avg = silhouette_score(X, cluster_labels)
    print("For n_clusters =", n_clusters,
          "The average silhouette_score is :", silhouette_avg)

    # Compute the silhouette scores for each sample
    sample_silhouette_values = silhouette_samples(X, cluster_labels)

    y_lower = 10
    for i in range(n_clusters):
        # Aggregate the silhouette scores for samples belonging to
        # cluster i, and sort them
        ith_cluster_silhouette_values = \
            sample_silhouette_values[cluster_labels == i]

        ith_cluster_silhouette_values.sort()

        size_cluster_i = ith_cluster_silhouette_values.shape[0]
        y_upper = y_lower + size_cluster_i

        color = cm.nipy_spectral(float(i) / n_clusters)
        ax1.fill_betweenx(np.arange(y_lower, y_upper),
                          0, ith_cluster_silhouette_values,
                          facecolor=color, edgecolor=color, alpha=0.7)

        # Label the silhouette plots with their cluster numbers at the middle
        ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))

        # Compute the new y_lower for next plot
        y_lower = y_upper + 10  # 10 for the 0 samples

    ax1.set_title("Diagrama de silueta por cada grupo.")
    ax1.set_xlabel("Coeficiente de silueta")
    ax1.set_ylabel("Grupos")

    # The vertical line for average silhouette score of all the values
    ax1.axvline(x=silhouette_avg, color="red", linestyle="--")

    ax1.set_yticks([])  # Clear the yaxis labels / ticks

    # 2nd Plot showing the actual clusters formed
    colors = cm.nipy_spectral(cluster_labels.astype(float) / n_clusters)
    ax2.scatter(X[:, 0], X[:, 1], marker='.', s=30, lw=0, alpha=0.7,
                c=colors, edgecolor='k')

    ax2.set_title("Subespacio CP1-CP2")
    ax2.set_xlabel("CP1 (ICE principal)")
    ax2.set_ylabel("CP2 (segundo ICE)")

    plt.suptitle(("Análisis de silueta con Average Linkage (7 indicadores base), "
                  "número de estratos = %d, " % n_clusters +
                 "score de silueta promedio = %1.3f" % silhouette_avg),
                 fontsize=14, fontweight='bold')
    #plt.savefig('Pyplot_silhouette_score_ind7_average_'+str(n_clusters)+'.png',dpi=600)

For n_clusters = 2 The average silhouette_score is : 0.7427739664649179
For n_clusters = 3 The average silhouette_score is : 0.7136632889400808
For n_clusters = 4 The average silhouette_score is : 0.6454909112785183
For n_clusters = 5 The average silhouette_score is : 0.6106028022406746
For n_clusters = 6 The average silhouette_score is : 0.5253221845133971
For n_clusters = 7 The average silhouette_score is : 0.5220856618240213

Dendrograma

import scipy.cluster.hierarchy as shc
plt.figure(figsize=(10,7))  
plt.title("Dendrograma, método de average linkage")  
dend = shc.dendrogram(shc.linkage(X, method='average'))
#plt.savefig('Pyplot_dendrograma_ind7_average.png',dpi=600)

Análisis de silueta, dendograma y configuración óptima de grupos con single linkage

Cálculo de los silhouette_scores para diferentes números de clusters en single linkage

from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import silhouette_samples, silhouette_score
import matplotlib.cm as cm

#Datos en el espacio de las componentes principales
X=x_pca

#numero de clusters/grupos que serán analizados
range_n_clusters = [2, 3, 4, 5,6,7]

for n_clusters in range_n_clusters:
    # Create a subplot with 1 row and 2 columns
    fig, (ax1, ax2) = plt.subplots(1, 2)
    fig.set_size_inches(18, 7)

    # The 1st subplot is the silhouette plot
    # The silhouette coefficient can range from -1, 1 
    ax1.set_xlim([-1, 1])
    # The (n_clusters+1)*10 is for inserting blank space between silhouette
    # plots of individual clusters, to demarcate them clearly.
    ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])

    # Initialize the clusterer with n_clusters value and a random generator
    # seed of 0 for reproducibility.
    clusterer = AgglomerativeClustering(
        n_clusters=n_clusters, linkage='single')
    cluster_labels = clusterer.fit_predict(X)

    # The silhouette_score gives the average value for all the samples.
    # This gives a perspective into the density and separation of the formed clusters
    silhouette_avg = silhouette_score(X, cluster_labels)
    print("For n_clusters =", n_clusters,
          "The average silhouette_score is :", silhouette_avg)

    # Compute the silhouette scores for each sample
    sample_silhouette_values = silhouette_samples(X, cluster_labels)

    y_lower = 10
    for i in range(n_clusters):
        # Aggregate the silhouette scores for samples belonging to
        # cluster i, and sort them
        ith_cluster_silhouette_values = \
            sample_silhouette_values[cluster_labels == i]

        ith_cluster_silhouette_values.sort()

        size_cluster_i = ith_cluster_silhouette_values.shape[0]
        y_upper = y_lower + size_cluster_i

        color = cm.nipy_spectral(float(i) / n_clusters)
        ax1.fill_betweenx(np.arange(y_lower, y_upper),
                          0, ith_cluster_silhouette_values,
                          facecolor=color, edgecolor=color, alpha=0.7)

        # Label the silhouette plots with their cluster numbers at the middle
        ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))

        # Compute the new y_lower for next plot
        y_lower = y_upper + 10  # 10 for the 0 samples

    ax1.set_title("Diagrama de silueta por cada grupo.")
    ax1.set_xlabel("Coeficiente de silueta")
    ax1.set_ylabel("Grupos")

    # The vertical line for average silhouette score of all the values
    ax1.axvline(x=silhouette_avg, color="red", linestyle="--")

    ax1.set_yticks([])  # Clear the yaxis labels / ticks

    # 2nd Plot showing the actual clusters formed
    colors = cm.nipy_spectral(cluster_labels.astype(float) / n_clusters)
    ax2.scatter(X[:, 0], X[:, 1], marker='.', s=30, lw=0, alpha=0.7,
                c=colors, edgecolor='k')

    ax2.set_title("Subespacio CP1-CP2")
    ax2.set_xlabel("CP1 (ICE principal)")
    ax2.set_ylabel("CP2 (segundo ICE)")

    plt.suptitle(("Análisis de silueta con Single Linkage (7 indicadores base), "
                  "número de estratos = %d, " % n_clusters +
                 "score de silueta promedio = %1.3f" % silhouette_avg),
                 fontsize=14, fontweight='bold')
    #plt.savefig('Pyplot_silhouette_score_ind7_single_'+str(n_clusters)+'.png',dpi=600)

For n_clusters = 2 The average silhouette_score is : 0.7479028873045614
For n_clusters = 3 The average silhouette_score is : 0.7077506062167823
For n_clusters = 4 The average silhouette_score is : 0.6288977274847589
For n_clusters = 5 The average silhouette_score is : 0.6200655258181543
For n_clusters = 6 The average silhouette_score is : 0.6256720086125872
For n_clusters = 7 The average silhouette_score is : 0.5759449091297045

Dendrograma

import scipy.cluster.hierarchy as shc
plt.figure(figsize=(10,7))  
plt.title("Dendrograma, método de single linkage")  
dend = shc.dendrogram(shc.linkage(X, method='single'))
#plt.savefig('Pyplot_dendrograma_ind7_single.png',dpi=600)

Escenario 1: Estratificacion con Ward (4 clusters, 3 CP, 7 indicadores)

Se utiliza la última iteración guardada de Ward para asegurar la reproducibilidad de los resultados

df_pca_results_ind7_ward.head()

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	CP1	CP2	CP3	CP4	CP5	CP6	CP7	Estrato
0	1	AGUASCALIENTES	10	SAN FRANCISCO DE LOS ROMO	-0.290827	1.304370	-0.470542	1.249692	-0.985389	0.343395	-0.883019	G1
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	-0.215022	-0.976397	-0.423919	1.386989	-1.260267	-0.625396	0.304296	G2
2	1	AGUASCALIENTES	9	TEPEZALA	-0.597435	0.071978	-0.716279	1.371428	-0.870282	-0.309656	-0.647832	G1
3	1	AGUASCALIENTES	11	EL LLANO	-0.382682	-0.278937	-0.606074	1.675445	-0.996878	-0.462573	-0.289389	G2
4	1	AGUASCALIENTES	2	ASIENTOS	-0.428400	-0.232114	-0.320612	1.533638	-0.432850	-0.508456	-0.335265	G2

Gráfica en el espacio CP1, CP2

#Scatter plot (seaborn) CP1 vs CP2 (con clustering k-means)

plt.figure(figsize=(16, 10), dpi= 200, facecolor='w', edgecolor='k')
sns.scatterplot(x="CP1", y="CP2", 
                data=df_pca_results_ind7_ward, 
                hue="Estrato",
                palette=color_dict)
# Set x-axis label
plt.xlabel('CP1 (ICE principal)')
# Set y-axis label
plt.ylabel('CP2 (segundo ICE)')
plt.title('Proyección en CP1-CP2 con estratificación (método de Ward)')

#plt.savefig('Pyplot_PCA_Ward_ind7.png',dpi=600)

Gráfica en el espacio CP1, CP2, CP3

from matplotlib import pyplot
from mpl_toolkits.mplot3d import Axes3D
from matplotlib.lines import Line2D

#para definir los colores
colores = df_pca_results_ind7_ward['Estrato'].apply(lambda x: color_dict[x])

#figura
fig = pyplot.figure(figsize=(16, 10), dpi=600, facecolor='w', edgecolor='k')
ax = Axes3D(fig)
ax.scatter(df_pca_results_ind7_ward['CP1'], df_pca_results_ind7_ward['CP2'], df_pca_results_ind7_ward['CP3'], c=colores,s=15)
# ax.set_xlim3d(-10, 40)
ax.set_ylim3d(-7.5,7.5)
ax.set_zlim3d(-2,8)
ax.set_xlabel('CP1 (ICE principal)')
ax.set_ylabel('CP2 (segundo ICE)')
ax.set_zlabel('CP3 (tercer ICE)')
ax.view_init(elev, azim)
ax.legend(title='Estratos', handles=legend_elements_5, loc='center right',fontsize='x-large')
ax.set_title('Proyección en CP1-CP2-CP3 con estratificación (método de Ward)')

#plt.savefig('Pyplot_PCA_3d_Ward_ind7.png',dpi=600)
pyplot.show()

Gráfica 3d animada

#3-d plot con plotly
fig = px.scatter_3d(df_pca_results_ind7_ward, x='CP1', y='CP2', z='CP3',
              color='Estrato',
                    color_discrete_map={
                  'G0':'blue',
                  'G1':'green',
                  'G2': 'orange',
                  'G3': 'grey',
                  'G4': 'red',
                  'G5': 'magenta'})
fig.update_traces(mode='markers', marker_line_width=1, marker_size=3)
fig.update_layout(scene = dict(
                    xaxis_title='CP1 (ICE principal)',
                    yaxis_title='CP2 (segundo ICE)',
                    zaxis_title='CP3 (tercer ICE)'),
                 legend_orientation="h")
fig.show()

Análisis estadístico de los ICE y la estratificacion

Escenario 1: Ward

#unimos los resultados de PCA_ward con los metadatos y los 7 indicadores de los municipios
df_indi_pca_ward_ind7=pd.merge(datos_mun_limpio, df_pca_results_ind7_ward, on=['EDO','NOMBRE_ESTADO','MUN','NOMBRE_MUNICIPIO'], how='inner')

#renombramos algunas columnas_pca
df_indi_pca_ward_ind7=df_indi_pca_ward_ind7.rename(columns={"CP1": "CP1 (ICE principal)", "CP2": "CP2 (segundo ICE)", "CP3": "CP3 (tercer ICE)"})

df_indi_pca_ward_ind7

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	NUM_SECC_RUR	LNE	Densidad_LNE	Razon_LNE_PE	TC_LNE_2019	Coef_Var_LNE	Var_Prop_LNE	Autocorr_LNE	CP1 (ICE principal)	CP2 (segundo ICE)	CP3 (tercer ICE)	CP4	CP5	CP6	CP7	Estrato
0	1	AGUASCALIENTES	10	SAN FRANCISCO DE LOS ROMO	3	34917.0	3566.024547	0.989272	0.037498	0.008903	0.017733	0.970914	-0.290827	1.304370	-0.470542	1.249692	-0.985389	0.343395	-0.883019	G1
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	4	6914.0	46.444421	0.991803	0.000289	0.002356	0.008887	0.951687	-0.215022	-0.976397	-0.423919	1.386989	-1.260267	-0.625396	0.304296	G2
2	1	AGUASCALIENTES	9	TEPEZALA	6	15545.0	2478.675412	0.990635	0.005108	0.005070	0.011921	0.832575	-0.597435	0.071978	-0.716279	1.371428	-0.870282	-0.309656	-0.647832	G1
3	1	AGUASCALIENTES	11	EL LLANO	6	15067.0	1505.749552	0.991016	-0.003373	0.004768	0.014232	0.962703	-0.382682	-0.278937	-0.606074	1.675445	-0.996878	-0.462573	-0.289389	G2
4	1	AGUASCALIENTES	2	ASIENTOS	11	34222.0	1484.722551	0.989080	0.001874	0.004824	0.015397	0.837235	-0.428400	-0.232114	-0.320612	1.533638	-0.432850	-0.508456	-0.335265	G2
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2451	32	ZACATECAS	45	TEPECHITLAN	13	6959.0	36.438707	0.983523	0.066677	0.001991	0.021905	0.577314	0.336526	-0.414118	0.879856	-0.236011	-0.054547	-0.515434	0.024942	G0
2452	32	ZACATECAS	14	GENERAL FRANCISCO R. MURGUIA	32	17372.0	14.595272	0.986450	0.082705	0.002967	0.026072	0.911464	-0.052455	-0.693651	2.394801	0.574858	0.201790	-0.068696	-0.577988	G0
2453	32	ZACATECAS	44	TABASCO	16	12527.0	53.144872	0.983221	0.100307	0.005293	0.035416	0.928948	0.815847	0.275372	1.779525	0.018950	-0.257829	0.094995	-0.099322	G0
2454	32	ZACATECAS	9	CHALCHIHUITES	10	8222.0	599.837116	0.982861	0.075052	0.003692	0.024771	0.819352	0.562309	0.233027	0.954854	-0.056961	-0.390447	-0.419723	-0.098396	G0
2455	32	ZACATECAS	11	TRINIDAD GARCIA DE LA CADENA	3	2779.0	15.815887	0.972699	0.113381	0.002873	0.041745	-0.273110	1.391414	0.517387	0.405960	-2.500965	1.204864	-0.454846	0.366341	G0

2456 rows × 20 columns

Gráficas (pair-plot, box-plots, violin-plots)

#sólo tomamos los indicadores y grupos
df_indi_pca_ward_ind7_f = df_indi_pca_ward_ind7[['NUM_SECC_RUR','LNE', 'Densidad_LNE', 'Razon_LNE_PE', 'TC_LNE_2019', 'Coef_Var_LNE', 'Autocorr_LNE', 'Estrato']]
#pair-plot (colores en funcion de grupo)
corr3=sns.pairplot(df_indi_pca_ward_ind7_f, 
                   hue="Estrato",
                   palette=color_dict)
corr3.fig.suptitle("Correlograma con estratificación (método de Ward)", y=1.01)

#guardar la pair-plot
corr3.savefig('Pyplot_correlogram_estratificacion_ind7.png',dpi=600)

Plotgrid de gráficas de caja para cada indicador e índice de complejidad electoral (por grupo)

sns.set(style="whitegrid")
fig, axes = plt.subplots(4, 2, figsize=(16, 20), dpi= 200)

ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["NUM_SECC_RUR"], palette=color_dict, orient='v', 
    ax=axes[0, 0])
ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["LNE"], palette=color_dict, orient='v', 
    ax=axes[0, 1])
ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Densidad_LNE"], palette=color_dict, orient='v', 
    ax=axes[1, 0])
ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Razon_LNE_PE"], palette=color_dict, orient='v', 
    ax=axes[1, 1])
ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["TC_LNE_2019"], palette=color_dict, orient='v', 
    ax=axes[2, 0])
ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Coef_Var_LNE"], palette=color_dict, orient='v', 
    ax=axes[2, 1])
ax = sns.boxplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Autocorr_LNE"], palette=color_dict, orient='v', 
    ax=axes[3, 0])

fig.delaxes(axes[3][1]) #borrar subplot sin usar
plt.subplots_adjust(hspace=0.3) #dar mas espacio vertical
#se guarda la gráfica
#plt.savefig('Pyplot_box-plots-grid_estratificacion_ind7.png',dpi=600)
plt.show()

Gráfica de violín (por grupos, para el indicador "CP1")

plt.figure(figsize=(16, 10), dpi= 300, facecolor='w', edgecolor='k')
sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["CP1 (ICE principal)"], palette=color_dict);
#se guarda la gráfica
#plt.savefig('Pyplot_violin_estratificacion_ind7.png',dpi=600)
plt.show()

sns.set(style="whitegrid")

fig, axes = plt.subplots(3, 1, figsize=(16, 20), dpi= 200)
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["CP1 (ICE principal)"], palette=color_dict, orient='v', 
    ax=axes[0])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["CP2 (segundo ICE)"], palette=color_dict, orient='v', 
    ax=axes[1])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["CP3 (tercer ICE)"], palette=color_dict, orient='v', 
    ax=axes[2])

#fig.delaxes(axes[3][1]) #borrar subplot sin usar
plt.subplots_adjust(hspace=0.3) #dar mas espacio vertical

#se guarda la gráfica
#plt.savefig('Pyplot_violin-plot-grid-CP_estratificacion_ind7.png',dpi=600)
plt.show()

Plotgrid de gráficas de violín para cada indicador e índice de complejidad electoral (por grupo)

sns.set(style="whitegrid")

fig, axes = plt.subplots(4, 2, figsize=(16, 20), dpi= 200)
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["NUM_SECC_RUR"], palette=color_dict, orient='v', 
    ax=axes[0, 0])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["LNE"], palette=color_dict, orient='v', 
    ax=axes[0, 1])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Densidad_LNE"], palette=color_dict, orient='v', 
    ax=axes[1, 0])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Razon_LNE_PE"], palette=color_dict, orient='v', 
    ax=axes[1, 1])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["TC_LNE_2019"], palette=color_dict, orient='v', 
    ax=axes[2, 0])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Coef_Var_LNE"], palette=color_dict, orient='v', 
    ax=axes[2, 1])
ax = sns.violinplot(x=df_indi_pca_ward_ind7["Estrato"], y=df_indi_pca_ward_ind7["Autocorr_LNE"], palette=color_dict, orient='v', 
    ax=axes[3, 0])

fig.delaxes(axes[3][1]) #borrar subplot sin usar
plt.subplots_adjust(hspace=0.3) #dar mas espacio vertical

#se guarda la gráfica
#plt.savefig('Pyplot_violin-plot-grid_estratificacion_ind7.png',dpi=600)
plt.show()

Identificación de los valores/municipios atípicos (7 indicadores)

Etiquetamos los municipios considerados como atípicos en los ICE (considerando la separación por grupos de la tipología, y utilizando la regla intercuartil):

#ejemplo:
df_indi_pca_ward_ind7['CP1 (ICE principal)'].quantile(0.25)

-0.4290536450983774

def IQsR_clusters(x):
    x_0 = x['Estrato'] == 'G0'
    x_g0 = x[x_0]
    Q1_g0 = x_g0['CP1 (ICE principal)'].quantile(0.25)
    Q3_g0 = x_g0['CP1 (ICE principal)'].quantile(0.75)
    IQR_pc1_g0 = Q3_g0 - Q1_g0
    Q1_pc2_g0 =  x_g0['CP2 (segundo ICE)'].quantile(0.25)
    Q3_pc2_g0 = x_g0['CP2 (segundo ICE)'].quantile(0.75)
    IQR_pc2_g0 = Q3_pc2_g0 - Q1_pc2_g0
    Q1_pc3_g0 =  x_g0['CP3 (tercer ICE)'].quantile(0.25)
    Q3_pc3_g0 = x_g0['CP3 (tercer ICE)'].quantile(0.75)
    IQR_pc3_g0 = Q3_pc3_g0 - Q1_pc3_g0
    return IQR_pc1_g0+IQR_pc1_g0;

IQsR_clusters(df_indi_pca_ward_ind7)

1.8897824926795055

Función para calcular el Q1, Q3 e IQR (calculado por grupos de cluster_ward), y posteriormente clasificar los municipios en tipicos o atipicos en el ICE principal.

def IQR_clusters(x):
    x_0 = df_indi_pca_ward_ind7['Estrato'] == 'G0'
    x_g0 = df_indi_pca_ward_ind7[x_0]
    Q1_g0 = x_g0['CP1 (ICE principal)'].quantile(0.25)
    Q3_g0 = x_g0['CP1 (ICE principal)'].quantile(0.75)
    IQR_pc1_g0 = Q3_g0 - Q1_g0
    Q1_pc2_g0 =  x_g0['CP2 (segundo ICE)'].quantile(0.25)
    Q3_pc2_g0 = x_g0['CP2 (segundo ICE)'].quantile(0.75)
    IQR_pc2_g0 = Q3_pc2_g0 - Q1_pc2_g0
    Q1_pc3_g0 =  x_g0['CP3 (tercer ICE)'].quantile(0.25)
    Q3_pc3_g0 = x_g0['CP3 (tercer ICE)'].quantile(0.75)
    IQR_pc3_g0 = Q3_pc3_g0 - Q1_pc3_g0
    
    x_1 = df_indi_pca_ward_ind7['Estrato'] == 'G1'
    x_g1 = df_indi_pca_ward_ind7[x_1]
    Q1_g1 = x_g1['CP1 (ICE principal)'].quantile(0.25)
    Q3_g1 = x_g1['CP1 (ICE principal)'].quantile(0.75)
    IQR_pc1_g1 = Q3_g1 - Q1_g1
    Q1_pc2_g1 =  x_g1['CP2 (segundo ICE)'].quantile(0.25)
    Q3_pc2_g1 = x_g1['CP2 (segundo ICE)'].quantile(0.75)
    IQR_pc2_g1 = Q3_pc2_g1 - Q1_pc2_g1
    Q1_pc3_g1 =  x_g1['CP3 (tercer ICE)'].quantile(0.25)
    Q3_pc3_g1 = x_g1['CP3 (tercer ICE)'].quantile(0.75)
    IQR_pc3_g1 = Q3_pc3_g1 - Q1_pc3_g1
    
    x_2 = df_indi_pca_ward_ind7['Estrato'] == 'G2'
    x_g2 = df_indi_pca_ward_ind7[x_2]
    Q1_g2 = x_g2['CP1 (ICE principal)'].quantile(0.25)
    Q3_g2 = x_g2['CP1 (ICE principal)'].quantile(0.75)
    IQR_pc1_g2 = Q3_g2 - Q1_g2
    Q1_pc2_g2 =  x_g2['CP2 (segundo ICE)'].quantile(0.25)
    Q3_pc2_g2 = x_g2['CP2 (segundo ICE)'].quantile(0.75)
    IQR_pc2_g2 = Q3_pc2_g2 - Q1_pc2_g2
    Q1_pc3_g2 =  x_g2['CP3 (tercer ICE)'].quantile(0.25)
    Q3_pc3_g2 = x_g2['CP3 (tercer ICE)'].quantile(0.75)
    IQR_pc3_g2 = Q3_pc3_g2 - Q1_pc3_g2
    
    x_3 = df_indi_pca_ward_ind7['Estrato'] == 'G3'
    x_g3 = df_indi_pca_ward_ind7[x_3]
    Q1_g3 = x_g3['CP1 (ICE principal)'].quantile(0.25)
    Q3_g3 = x_g3['CP1 (ICE principal)'].quantile(0.75)
    IQR_pc1_g3 = Q3_g3 - Q1_g3
    Q1_pc2_g3 =  x_g3['CP2 (segundo ICE)'].quantile(0.25)
    Q3_pc2_g3 = x_g3['CP2 (segundo ICE)'].quantile(0.75)
    IQR_pc2_g3 = Q3_pc2_g3 - Q1_pc2_g3
    Q1_pc3_g3 =  x_g3['CP3 (tercer ICE)'].quantile(0.25)
    Q3_pc3_g3 = x_g3['CP3 (tercer ICE)'].quantile(0.75)
    IQR_pc3_g3 = Q3_pc3_g3 - Q1_pc3_g3
    
    x_4 = df_indi_pca_ward_ind7['Estrato'] == 'G4'
    x_g4 = df_indi_pca_ward_ind7[x_4]
    Q1_g4 = x_g4['CP1 (ICE principal)'].quantile(0.25)
    Q3_g4 = x_g4['CP1 (ICE principal)'].quantile(0.75)
    IQR_pc1_g4 = Q3_g4 - Q1_g4
    Q1_pc2_g4 =  x_g4['CP2 (segundo ICE)'].quantile(0.25)
    Q3_pc2_g4 = x_g4['CP2 (segundo ICE)'].quantile(0.75)
    IQR_pc2_g4 = Q3_pc2_g4 - Q1_pc2_g4
    Q1_pc3_g4 =  x_g4['CP3 (tercer ICE)'].quantile(0.25)
    Q3_pc3_g4 = x_g4['CP3 (tercer ICE)'].quantile(0.75)
    IQR_pc3_g4 = Q3_pc3_g4 - Q1_pc3_g4
    
#     x_5 = df_indi_pca_ward_ind7['Estrato'] == 'G5'
#     x_g5 = df_indi_pca_ward_ind7[x_5]
#     Q1_g5 = x_g5['CP1 (ICE principal)'].quantile(0.25)
#     Q3_g5 = x_g5['CP1 (ICE principal)'].quantile(0.75)
#     IQR_pc1_g5 = Q3_g5 - Q1_g5
#     Q1_pc2_g5 =  x_g5['CP2 (segundo ICE)'].quantile(0.25)
#     Q3_pc2_g5 = x_g5['CP2 (segundo ICE)'].quantile(0.75)
#     IQR_pc2_g5 = Q3_pc2_g5 - Q1_pc2_g5
#     Q1_pc3_g5 =  x_g5['CP3 (tercer ICE)'].quantile(0.25)
#     Q3_pc3_g5 = x_g5['CP3 (tercer ICE)'].quantile(0.75)
#     IQR_pc3_g5 = Q3_pc3_g5 - Q1_pc3_g5
    
    if (x['Estrato'] == 'G0'):
        if (x['CP1 (ICE principal)'] < Q1_g0-1.5*IQR_pc1_g0) or (x['CP1 (ICE principal)'] > Q3_g0+1.5*IQR_pc1_g0):
            return 'Atípica en ICE principal'
        elif (x['CP2 (segundo ICE)'] < Q1_pc2_g0-1.5*IQR_pc2_g0) or (x['CP2 (segundo ICE)'] > Q3_pc2_g0+1.5*IQR_pc2_g0):
            return 'Atípica en segundo ICE'
        elif (x['CP3 (tercer ICE)'] < Q1_pc3_g0-1.5*IQR_pc3_g0) or (x['CP3 (tercer ICE)'] > Q3_pc3_g0+1.5*IQR_pc3_g0):
            return 'Atípica en tercer ICE'
        else:
            return 'Tipica'
    elif (x['Estrato'] == 'G1'):
        if (x['CP1 (ICE principal)'] < Q1_g1-1.5*IQR_pc1_g1) or (x['CP1 (ICE principal)'] > Q3_g1+1.5*IQR_pc1_g1):
            return 'Atípica en ICE principal'
        elif (x['CP2 (segundo ICE)'] < Q1_pc2_g1-1.5*IQR_pc2_g1) or (x['CP2 (segundo ICE)'] > Q3_pc2_g1+1.5*IQR_pc2_g1):
            return 'Atípica en segundo ICE'
        elif (x['CP3 (tercer ICE)'] < Q1_pc3_g1-1.5*IQR_pc3_g1) or (x['CP3 (tercer ICE)'] > Q3_pc3_g1+1.5*IQR_pc3_g1):
            return 'Atípica en tercer ICE'
        else:
            return 'Tipica'
    elif (x['Estrato'] == 'G2'):
        if (x['CP1 (ICE principal)'] < Q1_g2-1.5*IQR_pc1_g2) or (x['CP1 (ICE principal)'] > Q3_g2+1.5*IQR_pc1_g2):
            return 'Atípica en ICE principal'
        elif (x['CP2 (segundo ICE)'] < Q1_pc2_g2-1.5*IQR_pc2_g2) or (x['CP2 (segundo ICE)'] > Q3_pc2_g2+1.5*IQR_pc2_g2):
            return 'Atípica en segundo ICE'
        elif (x['CP3 (tercer ICE)'] < Q1_pc3_g2-1.5*IQR_pc3_g2) or (x['CP3 (tercer ICE)'] > Q3_pc3_g2+1.5*IQR_pc3_g2):
            return 'Atípica en tercer ICE'
        else:
            return 'Tipica'
    elif (x['Estrato'] == 'G3'):
        if (x['CP1 (ICE principal)'] < Q1_g3-1.5*IQR_pc1_g3) or (x['CP1 (ICE principal)'] > Q3_g3+1.5*IQR_pc1_g3):
            return 'Atípica en ICE principal'
        elif (x['CP2 (segundo ICE)'] < Q1_pc2_g3-1.5*IQR_pc2_g3) or (x['CP2 (segundo ICE)'] > Q3_pc2_g3+1.5*IQR_pc2_g3):
            return 'Atípica en segundo ICE'
        elif (x['CP3 (tercer ICE)'] < Q1_pc3_g3-1.5*IQR_pc3_g3) or (x['CP3 (tercer ICE)'] > Q3_pc3_g3+1.5*IQR_pc3_g3):
            return 'Atípica en tercer ICE'
        else:
            return 'Tipica'
    elif (x['Estrato'] == 'G4'):
        if (x['CP1 (ICE principal)'] < Q1_g4-1.5*IQR_pc1_g4) or (x['CP1 (ICE principal)'] > Q3_g4+1.5*IQR_pc1_g4):
            return 'Atípica en ICE principal'
        elif (x['CP2 (segundo ICE)'] < Q1_pc2_g4-1.5*IQR_pc2_g4) or (x['CP2 (segundo ICE)'] > Q3_pc2_g4+1.5*IQR_pc2_g4):
            return 'Atípica en segundo ICE'
        elif (x['CP3 (tercer ICE)'] < Q1_pc3_g4-1.5*IQR_pc3_g4) or (x['CP3 (tercer ICE)'] > Q3_pc3_g4+1.5*IQR_pc3_g4):
            return 'Atípica en tercer ICE'
        else:
            return 'Tipica'
#     elif (x['Estrato'] == 'G5'):
#         if (x['CP1 (ICE principal)'] < Q1_g5-1.5*IQR_pc1_g5) or (x['CP1 (ICE principal)'] > Q3_g5+1.5*IQR_pc1_g5):
#             return 'Atípica en ICE principal'
#         elif (x['CP2 (segundo ICE)'] < Q1_pc2_g5-1.5*IQR_pc2_g5) or (x['CP2 (segundo ICE)'] > Q3_pc2_g5+1.5*IQR_pc2_g5):
#             return 'Atípica en segundo ICE'
#         elif (x['CP3 (tercer ICE)'] < Q1_pc3_g5-1.5*IQR_pc3_g5) or (x['CP3 (tercer ICE)'] > Q3_pc3_g5+1.5*IQR_pc3_g5):
#             return 'Atípica en tercer ICE'
#         else:
#             return 'Tipica'
    else:
        return 'nan'

#se aplica la función al último df con los resultados:
df_indi_pca_ward_ind7['Atipicidad'] = df_indi_pca_ward_ind7.apply(IQR_clusters, axis = 1)

#se muestra el dataframe:
df_indi_pca_ward_ind7

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	NUM_SECC_RUR	LNE	Densidad_LNE	Razon_LNE_PE	TC_LNE_2019	Coef_Var_LNE	...	Autocorr_LNE	CP1 (ICE principal)	CP2 (segundo ICE)	CP3 (tercer ICE)	CP4	CP5	CP6	CP7	Estrato	Atipicidad
0	1	AGUASCALIENTES	10	SAN FRANCISCO DE LOS ROMO	3	34917.0	3566.024547	0.989272	0.037498	0.008903	...	0.970914	-0.290827	1.304370	-0.470542	1.249692	-0.985389	0.343395	-0.883019	G1	Tipica
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	4	6914.0	46.444421	0.991803	0.000289	0.002356	...	0.951687	-0.215022	-0.976397	-0.423919	1.386989	-1.260267	-0.625396	0.304296	G2	Tipica
2	1	AGUASCALIENTES	9	TEPEZALA	6	15545.0	2478.675412	0.990635	0.005108	0.005070	...	0.832575	-0.597435	0.071978	-0.716279	1.371428	-0.870282	-0.309656	-0.647832	G1	Tipica
3	1	AGUASCALIENTES	11	EL LLANO	6	15067.0	1505.749552	0.991016	-0.003373	0.004768	...	0.962703	-0.382682	-0.278937	-0.606074	1.675445	-0.996878	-0.462573	-0.289389	G2	Tipica
4	1	AGUASCALIENTES	2	ASIENTOS	11	34222.0	1484.722551	0.989080	0.001874	0.004824	...	0.837235	-0.428400	-0.232114	-0.320612	1.533638	-0.432850	-0.508456	-0.335265	G2	Tipica
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2451	32	ZACATECAS	45	TEPECHITLAN	13	6959.0	36.438707	0.983523	0.066677	0.001991	...	0.577314	0.336526	-0.414118	0.879856	-0.236011	-0.054547	-0.515434	0.024942	G0	Tipica
2452	32	ZACATECAS	14	GENERAL FRANCISCO R. MURGUIA	32	17372.0	14.595272	0.986450	0.082705	0.002967	...	0.911464	-0.052455	-0.693651	2.394801	0.574858	0.201790	-0.068696	-0.577988	G0	Atípica en tercer ICE
2453	32	ZACATECAS	44	TABASCO	16	12527.0	53.144872	0.983221	0.100307	0.005293	...	0.928948	0.815847	0.275372	1.779525	0.018950	-0.257829	0.094995	-0.099322	G0	Tipica
2454	32	ZACATECAS	9	CHALCHIHUITES	10	8222.0	599.837116	0.982861	0.075052	0.003692	...	0.819352	0.562309	0.233027	0.954854	-0.056961	-0.390447	-0.419723	-0.098396	G0	Tipica
2455	32	ZACATECAS	11	TRINIDAD GARCIA DE LA CADENA	3	2779.0	15.815887	0.972699	0.113381	0.002873	...	-0.273110	1.391414	0.517387	0.405960	-2.500965	1.204864	-0.454846	0.366341	G0	Tipica

2456 rows × 21 columns

#para verificar, imprimimos los valores únicos en la columna Atipicidad.
df_indi_pca_ward_ind7.Atipicidad.unique()

array(['Tipica', 'Atípica en ICE principal', 'Atípica en tercer ICE',
       'Atípica en segundo ICE'], dtype=object)

#se guardan los resultados en un csv
#df_indi_pca_ward_ind7.to_csv(r'Resultados_Mun_Indic_PCA_ward_Atip_ind7.csv', index = None)

Estadisticas

#resultados
df_indi_pca_ward_ind7 = pd.read_csv("E:/Google Drive/PREPRINTS_PONENCIAS/[Articulo]_Complejidad sistema geo-electoral_municipios/Code_Mun/3_Estratificacion/Resultados_Mun_Indic_PCA_ward_Atip_ind7.csv", encoding='utf-8')

#imprimimos la frecuencia de los valores únicos en Atipicidad
df_indi_pca_ward_ind7['Atipicidad'].value_counts()

Tipica                      2315
Atípica en ICE principal      64
Atípica en tercer ICE         45
Atípica en segundo ICE        32
Name: Atipicidad, dtype: int64

#imprimimos el número de municipios por grupo de la tipología.
df_indi_pca_ward_ind7['Estrato'].value_counts()

G2    1266
G0     815
G1     239
G4     116
G3      20
Name: Estrato, dtype: int64

#estadisticas tendencia central de los estratos
df_indi_pca_ward_ind7.groupby("Estrato")['LNE'].describe().reset_index()

	Estrato	count	mean	std	min	25%	50%	75%	max
0	G0	815.0	31832.851534	48633.809305	320.0	7511.00	16616.0	35772.00	695202.0
1	G1	239.0	140707.154812	178460.963273	327.0	16842.50	63802.0	204384.50	789602.0
2	G2	1266.0	10385.068720	13808.083953	152.0	2294.25	5892.5	13091.50	170909.0
3	G3	20.0	844279.050000	387284.987308	353944.0	474021.00	891754.0	1161875.75	1463720.0
4	G4	116.0	4418.646552	7729.344483	61.0	868.25	1854.0	4569.00	62803.0

#Estadisticas por estado de acuerdo a la estratificacion y al valor promedio de los ICE por estrato
apa = df_indi_pca_ward_ind7
apa['LNE_EDO'] = apa['LNE'].groupby(apa['NOMBRE_ESTADO']).transform('sum')
apa

stat_edo_estratos = apa.groupby(['NOMBRE_ESTADO','Estrato']).apply(lambda s: pd.Series({ 
    "NoMun": s["NOMBRE_MUNICIPIO"].count(), 
    "LNE_estrato": s["LNE"].sum(),
    "%LNE_estrato": s["LNE"].sum() / s["LNE_EDO"].mean(),
    "CP1_estrato": s["CP1 (ICE principal)"].mean(),
    "CP2_estrato": s["CP2 (segundo ICE)"].mean(),
    "CP3_estrato": s["CP3 (tercer ICE)"].mean(),
}))

stat_edo_estratos.reset_index().to_csv(r'Resultados_Mun_Edo-Stat_ind7.csv', index = None)
stat_edo_estratos

		NoMun	LNE_estrato	%LNE_estrato	CP1_estrato	CP2_estrato	CP3_estrato
NOMBRE_ESTADO	Estrato
AGUASCALIENTES	G1	7.0	861734.0	0.893580	-1.458196	0.908289	-0.713954
	G2	4.0	102627.0	0.106420	-0.373650	-0.574077	-0.346847
BAJA CALIFORNIA	G0	1.0	399621.0	0.142334	-2.395104	-0.022061	1.689324
	G1	3.0	972264.0	0.346295	-2.044081	1.125371	0.167323
	G3	1.0	1435734.0	0.511371	-9.091774	6.306996	1.612312
...	...	...	...	...	...	...	...
YUCATAN	G2	55.0	248197.0	0.156762	0.134916	-0.174617	-0.691249
	G4	1.0	9146.0	0.005777	1.702222	1.504063	1.518975
ZACATECAS	G0	43.0	831811.0	0.713527	0.167338	-0.216018	1.367701
	G1	3.0	256573.0	0.220088	-0.747096	0.531817	0.479643
	G2	12.0	77390.0	0.066385	0.205593	-0.518627	-0.236246

109 rows × 6 columns

#se ordenan y obtienen los 10 municipios por nivel de complejidad (descendente en CP1) 
ranking_mun_top10_ind7 = df_indi_pca_ward_ind7.sort_values(by='CP1 (ICE principal)', ascending=False).head(10)
#se guarda el ranking
#ranking_mun_top10_ind7.to_csv(r'Resultados_Mun_ranking-top10_ind7.csv', index = None)
ranking_mun_top10_ind7

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	NUM_SECC_RUR	LNE	Densidad_LNE	Razon_LNE_PE	TC_LNE_2019	Coef_Var_LNE	...	Autocorr_LNE	CP1 (ICE principal)	CP2 (segundo ICE)	CP3 (tercer ICE)	CP4	CP5	CP6	CP7	Estrato	Atipicidad
1502	20	OAXACA	472	SANTIAGO LALOPA	0	240.0	528.022755	0.947313	-0.250000	0.122907	...	0.589821	14.626957	11.542562	-8.691933	15.938933	11.055766	8.832353	-1.475366	G4	Atípica en ICE principal
1479	20	OAXACA	396	SANTA MAGDALENA JICOTLAN	0	61.0	37.305746	0.972984	0.033898	0.060861	...	0.539164	7.279769	5.791729	-2.360512	5.763452	3.972018	6.230660	-0.407567	G4	Atípica en ICE principal
218	8	CHIHUAHUA	66	URUACHI	10	4527.0	2.490662	0.946017	0.239595	0.018420	...	0.858316	5.714854	5.090749	3.613566	-2.578697	2.665680	-0.276434	-0.050782	G4	Atípica en ICE principal
1370	20	OAXACA	111	SAN BALTAZAR YATZACHI EL BAJO	2	477.0	34.518585	0.933182	0.060000	0.013315	...	0.305952	5.687124	4.205596	0.231008	-1.548074	4.494725	-4.109399	0.268599	G4	Atípica en ICE principal
1084	20	OAXACA	1	ABEJONES	0	747.0	531.676267	0.942882	0.131818	0.008108	...	0.793113	4.650944	4.005159	1.686696	-2.416453	2.364256	-3.300350	0.218580	G4	Tipica
1354	20	OAXACA	329	SAN PEDRO TIDAA	0	649.0	148.056274	0.969378	0.275049	0.024401	...	0.632966	4.629862	4.415049	2.851888	-2.062200	0.794104	3.537105	0.105549	G4	Atípica en segundo ICE
169	7	CHIAPAS	47	JITOTOL	5	13138.0	338.192820	0.957314	0.058151	0.024827	...	-0.056995	4.571080	3.438807	-0.754126	0.381572	3.876239	0.216013	-0.063830	G4	Tipica
1191	20	OAXACA	314	SAN PEDRO MARTIR QUIECHAPA	0	516.0	204.012472	0.957088	0.070539	0.024177	...	-0.710181	4.550628	3.167895	-1.404985	-0.730656	4.332651	0.650409	0.100386	G4	Tipica
1493	20	OAXACA	202	SAN JUAN JUQUILA VIJANOS	1	1322.0	1551.497965	0.956166	0.145581	0.020538	...	0.587084	4.510299	4.501003	1.002304	-1.013843	2.254619	0.081585	-0.381984	G4	Atípica en segundo ICE
1383	20	OAXACA	216	SAN JUAN TABAA	0	831.0	1138.040489	0.935902	0.183761	0.001318	...	-0.166667	4.228998	3.900017	1.882728	-5.318043	3.403119	-3.787944	0.117378	G4	Tipica

10 rows × 21 columns

#se ordenan y obtienen los 10 municipios por nivel de complejidad (ascendente en CP1) 
ranking_mun_low10_ind7 = df_indi_pca_ward_ind7.sort_values(by='CP1 (ICE principal)', ascending=True).head(10)
#se guarda el ranking
#ranking_mun_low10_ind7.to_csv(r'Resultados_Mun_ranking-low10_ind7.csv', index = None)
ranking_mun_low10_ind7

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	NUM_SECC_RUR	LNE	Densidad_LNE	Razon_LNE_PE	TC_LNE_2019	Coef_Var_LNE	...	Autocorr_LNE	CP1 (ICE principal)	CP2 (segundo ICE)	CP3 (tercer ICE)	CP4	CP5	CP6	CP7	Estrato	Atipicidad
266	9	CIUDAD DE MEXICO	7	IZTAPALAPA	0	1463720.0	17230.453735	0.988374	0.046250	0.001790	...	-0.163290	-12.048908	9.917034	-1.142757	0.222950	2.570151	-0.204829	2.836127	G3	Tipica
673	15	MEXICO	60	NEZAHUALCOYOTL	0	852207.0	24276.021766	0.988512	0.044559	0.001931	...	-0.190677	-10.845348	9.820307	-2.811423	-0.869985	0.886223	-0.460802	-3.430042	G3	Tipica
272	9	CIUDAD DE MEXICO	5	GUSTAVO A. MADERO	0	1032426.0	18322.312099	0.988378	0.045667	0.001492	...	-0.111745	-10.177991	8.638142	-1.761418	-0.347879	1.444153	-0.366782	-0.144833	G3	Tipica
694	15	MEXICO	34	ECATEPEC DE MORELOS	0	1234991.0	13617.495605	0.988559	0.051937	0.001834	...	-0.154305	-9.924663	7.953511	-0.898741	0.027527	2.146126	-0.031107	2.802598	G3	Tipica
577	14	JALISCO	41	GUADALAJARA	0	1235224.0	12447.353069	0.989382	0.058701	0.001694	...	0.005832	-9.625112	7.646354	-0.510415	0.163212	1.897478	0.051635	3.236309	G3	Tipica
12	2	BAJA CALIFORNIA	4	TIJUANA	26	1435734.0	6592.515390	0.988981	0.005467	0.005260	...	0.795296	-9.091774	6.306996	1.612312	3.407870	3.365193	-0.374414	5.642327	G3	Tipica
280	9	CIUDAD DE MEXICO	15	CUAUHTEMOC	0	477446.0	20572.568590	0.986857	0.043593	0.001655	...	-0.110782	-7.841070	7.371881	-2.738628	-1.174483	0.199567	-0.659612	-4.298435	G3	Tipica
1709	21	PUEBLA	115	PUEBLA	3	1279745.0	4191.464556	0.989225	-0.007137	0.002834	...	0.036879	-7.773966	4.819993	-0.336736	1.888368	2.845506	-0.334338	6.429381	G3	Tipica
265	9	CIUDAD DE MEXICO	6	IZTACALCO	0	353944.0	20177.362432	0.988164	0.044412	0.001687	...	-0.182919	-7.235253	6.662412	-2.918159	-1.348234	-0.088679	-0.465747	-4.892619	G3	Tipica
277	9	CIUDAD DE MEXICO	14	BENITO JUAREZ	0	366539.0	19859.099763	0.986669	0.043566	0.001597	...	-0.059689	-7.066362	6.765415	-2.738917	-1.242386	-0.077257	-0.709701	-4.697862	G3	Tipica

10 rows × 21 columns

#se obtienen los municipios de mayor complejidad (mayor CP1) para cada una de las 32 entidades
idx_rank_mun_up = df_indi_pca_ward_ind7.groupby(['EDO'])['CP1 (ICE principal)'].transform(max) == df_indi_pca_ward_ind7['CP1 (ICE principal)']
ranking_mun_edo_up = df_indi_pca_ward_ind7[idx_rank_mun_up]
#ranking_mun_edo_up.to_csv(r'Resultados_Mun_ranking-edo_ind7.csv', index = None)
ranking_mun_edo_up

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	NUM_SECC_RUR	LNE	Densidad_LNE	Razon_LNE_PE	TC_LNE_2019	Coef_Var_LNE	...	Autocorr_LNE	CP1 (ICE principal)	CP2 (segundo ICE)	CP3 (tercer ICE)	CP4	CP5	CP6	CP7	Estrato	Atipicidad
1	1	AGUASCALIENTES	8	SAN JOSE DE GRACIA	4	6914.0	46.444421	0.991803	0.000289	0.002356	...	0.951687	-0.215022	-0.976397	-0.423919	1.386989	-1.260267	-0.625396	0.304296	G2	Tipica
13	2	BAJA CALIFORNIA	3	TECATE	3	85233.0	1700.546549	0.988233	-0.012604	0.005274	...	0.891630	-0.506229	0.262616	-0.891563	1.694739	-0.599976	-0.748448	0.157863	G1	Tipica
20	3	BAJA CALIFORNIA SUR	2	MULEGE	17	40805.0	5.033360	0.982451	0.078356	0.002128	...	0.288921	0.144877	-0.386234	1.065829	-0.617600	0.601071	-0.281898	0.105354	G0	Tipica
26	4	CAMPECHE	7	PALIZADA	9	7006.0	5.649768	0.983489	0.035931	0.003208	...	0.807954	0.523708	-0.292755	0.359145	0.536766	-0.299406	-0.857918	0.151214	G0	Tipica
40	5	COAHUILA	34	SIERRA MOJADA	3	3555.0	0.609471	0.946643	0.040691	0.002872	...	-0.133939	3.325374	1.916581	-0.227076	-2.241184	3.456639	-4.114883	0.401892	G4	Tipica
73	6	COLIMA	7	IXTLAHUACAN	7	4986.0	26.897272	0.983974	0.076193	0.002262	...	0.336468	0.405714	-0.342483	0.466912	-0.751783	-0.109554	-0.175849	0.222039	G0	Tipica
169	7	CHIAPAS	47	JITOTOL	5	13138.0	338.192820	0.957314	0.058151	0.024827	...	-0.056995	4.571080	3.438807	-0.754126	0.381572	3.876239	0.216013	-0.063830	G4	Tipica
218	8	CHIHUAHUA	66	URUACHI	10	4527.0	2.490662	0.946017	0.239595	0.018420	...	0.858316	5.714854	5.090749	3.613566	-2.578697	2.665680	-0.276434	-0.050782	G4	Atípica en ICE principal
267	9	CIUDAD DE MEXICO	9	MILPA ALTA	0	107273.0	2559.563059	0.989102	0.060115	0.002708	...	-0.069564	-1.198345	0.375815	-0.801250	-0.831386	-0.123398	0.409315	0.128690	G1	Tipica
286	10	DURANGO	35	TAMAZULA	28	15952.0	3.905730	0.968309	0.003144	0.008171	...	0.905864	1.782883	0.602164	1.001933	1.565270	2.155724	-2.171754	-0.503852	G0	Tipica
320	11	GUANAJUATO	34	SANTA CATARINA	3	4449.0	77.068011	0.988469	0.050035	0.004282	...	0.510594	0.284550	-0.439303	-0.138367	0.152378	-0.624673	0.153012	0.298571	G2	Tipica
367	12	GUERRERO	76	ACATEPEC	25	19716.0	40.442197	0.966636	0.011284	0.014851	...	0.192578	2.419799	1.109030	0.090414	1.242367	3.393159	-0.878190	-0.476945	G4	Tipica
455	13	HIDALGO	20	ELOXOCHITLAN	7	2499.0	12.241952	0.986633	0.072532	0.021735	...	0.943708	2.344529	1.545803	0.224640	2.138979	0.147301	2.401457	-0.078911	G4	Tipica
644	14	JALISCO	63	MEZQUITIC	6	13993.0	13.184356	0.973196	0.197826	0.014500	...	0.771664	2.990376	2.644907	2.434741	-1.395087	0.387050	1.570184	0.158465	G4	Tipica
716	15	MEXICO	108	TONATICO	4	10591.0	547.294230	0.981441	0.077197	0.004813	...	0.988078	0.963677	0.684764	0.811691	0.006531	-0.694237	-0.496087	0.125341	G0	Tipica
809	16	MICHOACAN	53	MARCOS CASTELLANOS	6	9679.0	71.109115	0.980490	0.106298	0.008273	...	0.950592	1.558689	1.041453	1.282332	-0.102110	-0.383769	0.270721	0.181530	G0	Tipica
900	17	MORELOS	9	HUITZILAC	5	15191.0	319.822381	0.982867	0.093586	0.006952	...	0.960405	1.122759	0.797933	1.030621	0.040137	-0.652277	0.170220	0.171705	G0	Tipica
935	18	NAYARIT	6	HUAJICORI	17	7985.0	3.673322	0.979255	0.074987	0.003950	...	0.845038	0.916474	0.154518	1.465800	0.053868	0.233615	-0.749972	-0.119425	G0	Tipica
953	19	NUEVO LEON	10	CARMEN	1	28495.0	2675.641052	0.981918	0.201864	0.018327	...	0.948614	2.049458	3.530654	1.827794	-0.736692	-0.622370	2.865315	-0.635210	G4	Tipica
1502	20	OAXACA	472	SANTIAGO LALOPA	0	240.0	528.022755	0.947313	-0.250000	0.122907	...	0.589821	14.626957	11.542562	-8.691933	15.938933	11.055766	8.832353	-1.475366	G4	Atípica en ICE principal
1593	21	PUEBLA	53	CHIGMECATITLAN	0	1069.0	246.634063	0.974683	-0.028182	0.010509	...	0.941908	1.953946	0.912787	-1.176041	1.889721	0.542535	-1.613192	0.234723	G2	Atípica en ICE principal
1800	22	QUERETARO	10	EL MARQUES	8	126227.0	468.618879	0.985441	0.115217	0.007975	...	0.951613	0.423575	1.241313	1.567101	0.106310	-0.493816	0.853028	0.662012	G0	Tipica
1801	23	QUINTANA ROO	11	PUERTO MORELOS	5	19744.0	1524.130372	0.978663	0.018887	0.010990	...	0.936633	1.293162	1.340651	-0.383486	1.428023	0.263788	-0.605397	-0.305465	G4	Tipica
1818	24	SAN LUIS POTOSI	5	CARDENAS	6	14193.0	95.126992	0.979429	0.054536	0.005812	...	0.903552	1.196718	0.538423	0.538348	0.381273	-0.144105	-0.763815	0.232457	G0	Tipica
1877	25	SINALOA	8	ELOTA	33	31068.0	21.322511	0.986226	0.119608	0.005433	...	0.896890	0.231177	-0.181282	2.940518	0.260099	0.295994	0.698179	-0.564722	G0	Atípica en tercer ICE
1934	26	SONORA	9	BACERAC	1	1032.0	1.679872	0.958345	0.108485	0.004879	...	0.974038	3.180450	2.376082	1.505513	-1.509914	0.889418	-2.507346	0.404480	G4	Tipica
1971	27	TABASCO	7	EMILIANO ZAPATA	5	23000.0	90.071946	0.983943	0.059273	0.003535	...	0.418046	0.440311	-0.156021	0.140199	-0.286897	-0.103822	-0.253335	0.354478	G2	Tipica
2003	28	TAMAULIPAS	24	MIER	0	4072.0	570.127109	0.987154	-0.055001	0.008804	...	0.916659	0.582293	-0.196311	-1.768147	2.610977	-0.410929	-0.859801	0.139655	G2	Tipica
2040	29	TLAXCALA	19	TEPETITLA DE LARDIZABAL	0	14740.0	1038.462771	0.984461	0.077170	0.005820	...	0.858303	0.703292	0.769300	0.343651	-0.018693	-0.914185	0.020521	0.089190	G0	Tipica
2272	30	VERACRUZ	107	MIAHUATLAN	1	3286.0	133.502548	0.982124	0.136631	0.006907	...	0.937054	1.485893	1.123170	1.486327	-0.820376	-0.972043	0.607251	0.312508	G0	Tipica
2301	31	YUCATAN	13	CONKAL	2	9146.0	342.222929	0.982072	0.142108	0.009826	...	0.955260	1.702222	1.504063	1.518975	-0.557390	-0.785238	1.058126	0.194292	G4	Tipica
2455	32	ZACATECAS	11	TRINIDAD GARCIA DE LA CADENA	3	2779.0	15.815887	0.972699	0.113381	0.002873	...	-0.273110	1.391414	0.517387	0.405960	-2.500965	1.204864	-0.454846	0.366341	G0	Tipica

32 rows × 21 columns

#se obtienen los municipios de menor complejidad (menor CP1) para cada una de las 32 entidades
idx_rank_mun_low = df_indi_pca_ward_ind7.groupby(['EDO'])['CP1 (ICE principal)'].transform(min) == df_indi_pca_ward_ind7['CP1 (ICE principal)']
ranking_mun_edo_low = df_indi_pca_ward_ind7[idx_rank_mun_low]
ranking_mun_edo_low.to_csv(r'Resultados_Mun_ranking-edo_ind7.csv', index = None)
ranking_mun_edo_low

	EDO	NOMBRE_ESTADO	MUN	NOMBRE_MUNICIPIO	NUM_SECC_RUR	LNE	Densidad_LNE	Razon_LNE_PE	TC_LNE_2019	Coef_Var_LNE	...	Autocorr_LNE	CP1 (ICE principal)	CP2 (segundo ICE)	CP3 (tercer ICE)	CP4	CP5	CP6	CP7	Estrato	Atipicidad
10	1	AGUASCALIENTES	1	AGUASCALIENTES	9	648368.0	5716.694188	0.990718	0.000471	0.004334	...	0.707402	-4.842460	3.220714	-0.321650	1.973270	0.719098	-0.407201	1.850425	G1	Atípica en ICE principal
12	2	BAJA CALIFORNIA	4	TIJUANA	26	1435734.0	6592.515390	0.988981	0.005467	0.005260	...	0.795296	-9.091774	6.306996	1.612312	3.407870	3.365193	-0.374414	5.642327	G3	Tipica
17	3	BAJA CALIFORNIA SUR	3	LA PAZ	28	206629.0	35.135769	0.989726	0.059245	0.003103	...	0.844457	-1.281797	-0.285719	1.934086	1.171503	0.389934	0.073320	0.677012	G0	Tipica
30	4	CAMPECHE	1	CAMPECHE	22	205770.0	3469.451374	0.990081	0.046867	0.001688	...	0.083429	-2.523880	0.302166	0.347423	0.109901	0.824965	0.126470	-0.363945	G0	Atípica en ICE principal
38	5	COAHUILA	30	SALTILLO	37	589670.0	190.077309	0.988114	0.073456	0.005078	...	0.790444	-3.069760	1.333327	2.996571	1.737869	1.940414	0.430369	2.578019	G1	Atípica en tercer ICE
76	6	COLIMA	5	VILLA DE ALVAREZ	4	101527.0	4134.640462	0.988143	0.072057	0.002949	...	0.517021	-1.341737	1.232043	-0.050734	-0.390184	-0.693170	0.073707	-0.633430	G1	Tipica
82	7	CHIAPAS	102	TUXTLA GUTIERREZ	0	444865.0	2690.244836	0.989995	0.051072	0.002147	...	0.243378	-2.995578	1.676227	-0.241355	0.053017	0.210878	0.156620	2.103133	G1	Tipica
206	8	CHIHUAHUA	37	JUAREZ	7	1097905.0	3388.880560	0.987540	0.072989	0.003037	...	0.637788	-6.131007	4.781923	1.570313	1.096355	1.692949	0.096761	5.499845	G3	Tipica
266	9	CIUDAD DE MEXICO	7	IZTAPALAPA	0	1463720.0	17230.453735	0.988374	0.046250	0.001790	...	-0.163290	-12.048908	9.917034	-1.142757	0.222950	2.570151	-0.204829	2.836127	G3	Tipica
282	10	DURANGO	5	DURANGO	54	475936.0	118.699310	0.991086	-0.007104	0.002743	...	0.688533	-3.546346	-0.570966	2.465498	3.081441	2.586182	-0.460766	1.363380	G1	Atípica en tercer ICE
363	11	GUANAJUATO	20	LEON	42	1137504.0	2515.161550	0.989268	0.068865	0.002871	...	0.659043	-6.917116	3.611348	3.535886	2.202442	3.296766	0.264707	4.856366	G3	Atípica en tercer ICE
439	12	GUERRERO	1	ACAPULCO DE JUAREZ	53	573423.0	1608.985529	0.987577	0.024372	0.001458	...	-0.096877	-4.410904	0.211441	2.215130	1.483843	3.709577	-0.184286	1.465020	G1	Atípica en ICE principal
483	13	HIDALGO	47	PACHUCA DE SOTO	2	229776.0	2363.736249	0.987842	0.047169	0.002087	...	0.053102	-1.763480	0.731686	-0.566388	-0.392851	0.236081	-0.015475	0.873433	G1	Tipica
577	14	JALISCO	41	GUADALAJARA	0	1235224.0	12447.353069	0.989382	0.058701	0.001694	...	0.005832	-9.625112	7.646354	-0.510415	0.163212	1.897478	0.051635	3.236309	G3	Tipica
673	15	MEXICO	60	NEZAHUALCOYOTL	0	852207.0	24276.021766	0.988512	0.044559	0.001931	...	-0.190677	-10.845348	9.820307	-2.811423	-0.869985	0.886223	-0.460802	-3.430042	G3	Tipica
822	16	MICHOACAN	54	MORELIA	29	611330.0	1872.366121	0.988685	0.061094	0.002802	...	0.498226	-3.880074	1.660309	2.003848	1.155287	1.771904	0.148994	2.395030	G1	Atípica en ICE principal
921	17	MORELOS	7	CUERNAVACA	1	311797.0	3760.291575	0.988121	0.060722	0.002070	...	-0.059313	-2.548254	1.516294	-0.582133	-0.706657	0.369894	0.205295	0.884701	G1	Tipica
939	18	NAYARIT	17	TEPIC	29	306177.0	319.982758	0.990310	0.049626	0.002002	...	0.504744	-2.171892	-0.235904	1.644693	1.008794	0.997202	0.081940	1.151228	G0	Atípica en ICE principal
991	19	NUEVO LEON	40	MONTERREY	1	931301.0	5879.643516	0.988187	0.071643	0.002814	...	0.568237	-5.940125	4.982857	0.655171	0.567276	0.944911	0.057484	3.791331	G3	Tipica
1165	20	OAXACA	66	OAXACA DE JUAREZ	0	221359.0	3802.931586	0.989247	0.041567	0.001639	...	-0.108299	-2.292109	0.941625	-1.097722	-0.593270	0.143232	0.054879	0.365767	G1	Tipica
1709	21	PUEBLA	115	PUEBLA	3	1279745.0	4191.464556	0.989225	-0.007137	0.002834	...	0.036879	-7.773966	4.819993	-0.336736	1.888368	2.845506	-0.334338	6.429381	G3	Tipica
1787	22	QUERETARO	14	QUERETARO	9	751167.0	2655.664379	0.987832	0.066250	0.003540	...	0.620809	-4.222046	3.169263	1.191377	0.891589	1.072554	0.115356	3.619194	G1	Atípica en ICE principal
1810	23	QUINTANA ROO	1	BENITO JUAREZ	3	605897.0	354.439321	0.989799	0.014013	0.005354	...	0.824120	-2.836040	1.834995	0.245964	1.996022	0.504501	-0.118840	3.771238	G1	Tipica
1845	24	SAN LUIS POTOSI	28	SAN LUIS POTOSI	33	624076.0	1018.139096	0.988268	0.066702	0.003193	...	0.669123	-3.648952	1.531102	2.574101	1.428131	1.848761	0.152332	2.643056	G1	Atípica en tercer ICE
1871	25	SINALOA	6	CULIACAN	149	695202.0	185.975005	0.988409	0.070901	0.002458	...	0.472069	-6.095242	-1.572388	8.882531	4.099044	7.612246	0.406199	-0.597798	G0	Atípica en ICE principal
1943	26	SONORA	59	CAJEME	16	317070.0	4670.581652	0.989790	0.060655	0.002659	...	0.765522	-2.916484	1.818450	0.834171	0.646823	-0.055512	-0.062187	0.044011	G1	Tipica
1966	27	TABASCO	4	CENTRO	47	501305.0	475.575106	0.988887	0.038529	0.001276	...	-0.211009	-3.750439	-0.337137	2.044256	0.977916	3.268939	0.184637	1.662227	G1	Atípica en tercer ICE
2008	28	TAMAULIPAS	32	REYNOSA	11	513438.0	842.152918	0.991406	-0.006275	0.004378	...	0.928261	-2.905947	1.154410	0.292920	2.454415	0.420768	-0.407096	2.771679	G1	Tipica
2057	29	TLAXCALA	59	SAN LORENZO AXOCOMANITLA	0	3720.0	4478.230867	0.990818	0.059829	0.001515	...	0.000000	-1.423675	0.397924	-1.082502	-1.031450	-0.739568	0.303746	-1.186129	G1	Tipica
2227	30	VERACRUZ	192	VERACRUZ	5	457975.0	5449.345732	0.990370	0.051308	0.001646	...	-0.030497	-4.054253	2.251494	-0.526576	-0.268693	0.625233	0.220427	0.999603	G1	Atípica en ICE principal
2344	31	YUCATAN	50	MERIDA	12	701334.0	2505.786073	0.988765	0.069885	0.003536	...	0.662667	-4.038847	2.839929	1.390445	0.933232	0.960707	0.231446	3.271988	G1	Atípica en ICE principal
2409	32	ZACATECAS	10	FRESNILLO	103	163706.0	540.775731	0.989945	0.070702	0.003495	...	0.911365	-2.527961	-2.063262	6.059075	2.900712	3.632123	0.311764	-2.350407	G0	Atípica en ICE principal

32 rows × 21 columns

Referencias

• https://www.analyticsvidhya.com/blog/2018/08/dimensionality-reduction-techniques-python/
• https://plot.ly/python/v3/ipython-notebooks/principal-component-analysis/#pca-vs-lda
• https://www.machinelearningplus.com/plots/top-50-matplotlib-visualizations-the-master-plots-python
• https://towardsdatascience.com/visualising-high-dimensional-datasets-using-pca-and-t-sne-in-python-8ef87e7915b
• https://medium.com/@indreshbhattacharyya/feature-selection-categorical-feature-selection-boruta-light-gbm-chi-square-bf47e94e2558
• https://stackoverflow.com/questions/52472757/creating-a-boxplot-facetgrid-in-seaborn-for-python
• https://www.kdnuggets.com/2019/07/10-simple-hacks-speed-data-analysis-python.html
• https://www.datanovia.com/en/lessons/assessing-clustering-tendency/
• https://docs.scipy.org/doc/scipy/reference/generated/scipy.cluster.hierarchy.linkage.html
• https://scikit-learn.org/stable/modules/generated/sklearn.cluster.AgglomerativeClustering.html#sklearn.cluster.AgglomerativeClustering